interact with the Git data structures.
"""
+from __future__ import absolute_import
import errno, os, sys, zlib, time, subprocess, struct, stat, re, tempfile, glob
from collections import namedtuple
from itertools import islice
from numbers import Integral
-from bup import _helpers, hashsplit, path, midx, bloom, xstat
+from bup import _helpers, compat, hashsplit, path, midx, bloom, xstat
from bup.helpers import (Sha1, add_error, chunkyreader, debug1, debug2,
fdatasync,
hostname, localtime, log, merge_iter,
mmap_read, mmap_readwrite,
parse_num,
- progress, qprogress, stat_if_exists,
+ progress, qprogress, shstr, stat_if_exists,
unlink, username, userfullname,
utc_offset_str)
def _git_wait(cmd, p):
rv = p.wait()
if rv != 0:
- raise GitError('%s returned %d' % (cmd, rv))
+ raise GitError('%s returned %d' % (shstr(cmd), rv))
def _git_capture(argv):
p = subprocess.Popen(argv, stdout=subprocess.PIPE, preexec_fn = _gitenv())
def repo(sub = '', repo_dir=None):
"""Get the path to the git repository or one of its subdirectories."""
- global repodir
repo_dir = repo_dir or repodir
if not repo_dir:
raise GitError('You should call check_repo_or_die()')
# If there's a .git subdirectory, then the actual repo is in there.
gd = os.path.join(repo_dir, '.git')
if os.path.exists(gd):
- repodir = gd
+ repo_dir = gd
return os.path.join(repo_dir, sub)
if self.bloom is None and os.path.exists(bfull):
self.bloom = bloom.ShaBloom(bfull)
self.packs = list(set(d.values()))
- self.packs.sort(lambda x,y: -cmp(len(x),len(y)))
+ self.packs.sort(reverse=True, key=lambda x: len(x))
if self.bloom and self.bloom.valid() and len(self.bloom) >= len(self):
self.do_bloom = True
else:
return None
-def rev_list(ref, count=None, repo_dir=None):
- """Generate a list of reachable commits in reverse chronological order.
+def rev_list_invocation(ref_or_refs, count=None, format=None):
+ if isinstance(ref_or_refs, compat.str_type):
+ refs = (ref_or_refs,)
+ else:
+ refs = ref_or_refs
+ argv = ['git', 'rev-list']
+ if isinstance(count, Integral):
+ argv.extend(['-n', str(count)])
+ elif count:
+ raise ValueError('unexpected count argument %r' % count)
+
+ if format:
+ argv.append('--pretty=format:' + format)
+ for ref in refs:
+ assert not ref.startswith('-')
+ argv.append(ref)
+ argv.append('--')
+ return argv
+
- This generator walks through commits, from child to parent, that are
- reachable via the specified ref and yields a series of tuples of the form
- (date,hash).
+def rev_list(ref_or_refs, count=None, parse=None, format=None, repo_dir=None):
+ """Yield information about commits as per "git rev-list". If a format
+ is not provided, yield one hex hash at a time. If a format is
+ provided, pass it to rev-list and call parse(git_stdout) for each
+ commit with the stream positioned just after the rev-list "commit
+ HASH" header line. When a format is provided yield (oidx,
+ parse(git_stdout)) for each commit.
- If count is a non-zero integer, limit the number of commits to "count"
- objects.
"""
- assert(not ref.startswith('-'))
- opts = []
- if isinstance(count, Integral):
- opts += ['-n', str(count)]
- else:
- assert not count
- argv = ['git', 'rev-list', '--pretty=format:%at'] + opts + [ref, '--']
- p = subprocess.Popen(argv,
+ assert bool(parse) == bool(format)
+ p = subprocess.Popen(rev_list_invocation(ref_or_refs, count=count,
+ format=format),
preexec_fn = _gitenv(repo_dir),
stdout = subprocess.PIPE)
- commit = None
- for row in p.stdout:
- s = row.strip()
- if s.startswith('commit '):
- commit = s[7:].decode('hex')
- else:
- date = int(s)
- yield (date, commit)
+ if not format:
+ for line in p.stdout:
+ yield line.strip()
+ else:
+ line = p.stdout.readline()
+ while line:
+ s = line.strip()
+ if not s.startswith('commit '):
+ raise Exception('unexpected line ' + s)
+ yield s[7:], parse(p.stdout)
+ line = p.stdout.readline()
+
rv = p.wait() # not fatal
if rv:
raise GitError, 'git rev-list returned error %d' % rv
KeyError.__init__(self, 'object %r is missing' % oid.encode('hex'))
-WalkItem = namedtuple('WalkItem', ['id', 'type', 'mode',
+WalkItem = namedtuple('WalkItem', ['oid', 'type', 'mode',
'path', 'chunk_path', 'data'])
# The path is the mangled path, and if an item represents a fragment
# of a chunked file, the chunk_path will be the chunked subtree path
pending = [(oidx, [], [], None)]
while len(pending):
oidx, parent_path, chunk_path, mode = pending.pop()
+ oid = oidx.decode('hex')
if stop_at and stop_at(oidx):
continue
# If the object is a "regular file", then it's a leaf in
# the graph, so we can skip reading the data if the caller
# hasn't requested it.
- yield WalkItem(id=oidx, type='blob',
+ yield WalkItem(oid=oid, type='blob',
chunk_path=chunk_path, path=parent_path,
mode=mode,
data=None)
else:
data = ''.join(item_it)
- yield WalkItem(id=oidx, type=typ,
+ yield WalkItem(oid=oid, type=typ,
chunk_path=chunk_path, path=parent_path,
mode=mode,
data=(data if include_data else None))
projects / bup.git / blobdiff